Implications of genotype-by-environment interactions in maize(zea mays L.)variery selection using mother and baby trials

Abstract

Maize is the main staple food crop in the Southern Africa Development Community (SADC), which includes Zimbabwe and its productivity increases when improved varieties are planted. Development of such technologies is mainly by public and private breeding programs, which in most cases do this in managed stress environments such as research stations while they are deployed in farmers' fields, which have random stresses. This situation creates Genotype by Environment (GxE) interactions, which constitute one of the major complications in plant breeding and has been widely discussed, particularly in relation to their effect on the choice of selection environment(s) and indeed the whole selection process. Repeatability of aspects of GxE interactions is an important factor to be assessed in designing more efficient selection programs. CIMMYT with collaborating National Agricultural Research Systems (NARSs) in the SADC region have been using an evaluation scheme, Mother and Baby Trials (MBTs), to ensure that maize varieties are tested in target environments. In Zimbabwe this testing has been going on since 1999. Analysis of historical data from these multi-environment trials over years can be used to provide some clues on how to improve the testing strategy. Using sequential retrospective pattern analysis based on the SEQRJET package Versionl.l the historical data from Zimbabwe were analyzed to characterize environments, the tested genotypes and identify repeatable GxE interactions across years. The results showed that environments were classified into four groups based on their similarities. Differences among these groups were mainly due to SWHC, EVAPO and soil depth as major determinants of favourable soil water status for plant growth. Elevation and temperature were also important. High grain yield was associated with environment groups characterized by high SWHC, low EVAPO and deep soil profile. Genotypes were also classified into four groups. The type of variety underpinned differences among genotype groups. OPVs, ZM 421, ZM 423 and ZM 521, bred for stress tolerance were in the same group while the hybrids, which were not specifically bred for stress tolerance, were in other groups. Varying soil nitrogen levels effectively created differe-it environments among the locations used. Inclusion of environments representing important stress in the testing environments improved the effectiveness of the selection programme in identifying suitable varieties for the conditions characteristic of farmers' environment in Zimbabwe. Variety testing in Zimbabwe must be rationalized to cover 16 environments represent two locations in agro-ecological zone II, one location in zone III, five locations in zone IV and one locations in zone V.